Tumor antigens on neoplasms induced by chemical carcinogens and by DNA- and RNA-containing viruses: properties of the solubilized antigens

The role of tumor rejection antigens in host antitumor defense mechanisms

BACKGROUND

Normal cells undergo contact inhibition of growth when their surface molecules interact. Tumor cells, however, have undergone a mutation that prevents this arrest of growth upon contact inhibition and allows constant growth. Thus, growth inhibition fails to occur despite the interaction of surface molecules. In recent years a subgroup of these surface molecules has been of interest to cancer investigators. This subgroup has been termed the tumor rejection antigens (TRAs). As the name implies, these are specific to the tumor of origin and may direct the immune system of the host to target the tumor cells and kill them.

METHODS

A literature search was carried out on TRAs to ascertain the current thinking on the subject.

RESULTS

Initial studies of TRAs have revealed that some of them may be heat shock proteins (HSPs). In particular, grp96, a number of the HSP90 family, has been implicated. More recent studies, however, have shown that HSPs alone may not be immunogenic but may act as carrier proteins for tumor specific peptides.

CONCLUSION

Such findings have led to speculation that HSPs or their associated peptides may have a role in the diagnosis and/or treatment of specific cancers. Immunotherapy and bispecific antibodies in particular are areas in which HSPs may prove to be useful.

The gene coding for a major tumor rejection antigen of tumor P815 is identical to the normal gene of syngeneic DBA/2 mice

We showed previously that mouse mastocytoma P815 expresses several distinct antigens that are recognized by cytolytic T lymphocytes (CTL) of syngeneic DBA/2 mice. Antigens P815A and P815B are usually lost jointly and are targets for immune rejection responses in vivo. We used a cosmid library and a CTL stimulation assay to obtain transfectants expressing tumor rejection antigen P815A. From these transfectants we retrieved gene P1A which transferred the expression of both P815A and B. This gene is unrelated to three previously isolated genes coding for tum-antigens. It encodes a putative protein of 224 amino acids which contains two highly acidic domains showing homology with similar regions of nuclear proteins. The P1A gene expressed by tumor P815 is completely identical to the gene present in normal DBA/2 cells. Expression of the gene was tested by Northern blots. Cells from liver, spleen, and a number of mast cell lines were negative, but mast cell line L138.8A produced a high level of P1A message and was lysed by CTL directed against antigens P815A and B. We conclude that major tumor rejection antigens of P815 are encoded by a gene showing little or no expression in most normal cells of adult mice.

The tumor rejection antigen separated from Rous sarcoma virus-induced murine fibrosarcoma exhibits a molecular weight of approximately 60 kD but differs from functional pp60src

The tumor antigen capable of inducing tumor resistance (tumor rejection antigen; TRA) was obtained in a solubilized form by sodium dodecyl sulfate (SDS) extraction of plasma membrane fraction from Rous sarcoma virus (RSV)-induced CSA1M fibrosarcoma cells (BALB/c origin). Analyses by Sephacryl S-300 gel filtration and SDS-polyacrylamide gel electrophoresis revealed that TRA activity was recovered in the fraction with a molecular weight of approximately 60 kD. Unfractionated crude SDS-solubilized preparation contained gp70 as detected by rabbit anti-gp70 antiserum, whereas such reactivity was lost in the fraction exhibiting the molecular weight of about 60 kD. Since this fraction retained pp60src activity, the relation of TRA to pp60src was further investigated. pp60v-src was also obtained from the lysate of v-src-expressing yeast transformant. Immunization of BALB/c mice with such pp60v-src-containing lysate failed to induce any significant tumor protection. The above 60 kD fraction of CSA1M solubilized antigens was allowed to bind to Sepharose beads coupled with anti-pp60src monoclonal antibody and separated into the bead-bound and bead-unbound fractions. The bead-bound fraction that was recovered from pp60src-binding beads (pp60src-positive fraction) did not exhibit the TRA activity. In contrast, immunization with the fraction depleted of pp60src activity (bead-unbound fraction) resulted in potent tumor protection. These results indicate that the solubilized membranous component(s) of CSA1M with a molecular weight of approximately 60 kD, which is distinct from functional pp60src, functions as the TRA against RSV-induced CSA1M tumor cells.

Separation of the tumor rejection antigen of Rous sarcoma virus-induced murine fibrosarcoma

The tumor antigen capable of inducing tumor resistance (tumor rejection antigen; TRA) was separated and some of its physicochemical properties were characterized. Cytosol and plasma membrane fractions were separated from Rous sarcoma virus (RSV)-induced CSA1M tumor cells. Immunization with membrane but not cytosol fraction of these tumor cells together with complete Freund's adjuvant resulted in complete protection against subsequent challenge with viable CSA1M cells. The TRA activity contained in the membrane fraction was recovered in the sodium dodecyl sulfate (SDS)-solubilized fraction after the SDS-extraction of CSA1M membranes. This CSA1M SDS-solubilized preparation gave protection against syngeneic RSV-induced CSA9F tumor cells as well as the homologous tumor cell type, but failed to induce resistance to RSV-unrelated tumor cells. The membrane or SDS-solubilized fraction from RSV-unrelated tumor cells was unable to generate anti-CSA1M protective immunity. Physicochemical analyses have demonstrated that TRA activity in the SDS-solubilized fraction was completely abolished by treatment with proteinase K but was only marginally affected after treatment with glycosidase mixture. When the SDS-solubilized preparation was applied to a Sephacryl S-300 superfine column, TRA activity was recovered in the range of molecular weight of 50-90 kD. Further fractionation of this TRA-positive fraction by SDS-polyacrylamide gel electrophoresis revealed that the molecular size of TRA is 56-68 kD. These results indicate that membrane proteins which were isolated from CSA1M tumor cells and have a molecular size of about 60 kD are capable of inducing RSV-induced tumor-specific in vivo protective immunity.

Partial purification and characterization of a murine glioma-associated antigen defined by syngeneic rat monoclonal antibodies

A glioma-associated antigen was previously identified on an avian sarcoma virus-induced F-344 rat astrocytoma cell line S69-c15 by four rat monoclonal antibodies (7G4, 9F1, 10E3 and 10E7) produced after syngeneic immunization. Earlier data suggested all four antibodies reacted with a polypeptide-associated epitope. We report here that the antigen activity was detected in the supernatant of tumor homogenates and could pass through a 1000 Da molecular weight cut-off dialysis membrane, as determined by antibody binding inhibition in a cell surface radioimmunoassay. When the dialysate was fractionated by Bio-Gel P-2 chromatography, antibody inhibiting activity eluted in the range of 300-600 Da. A highly purified material was further isolated by ion exchange high pressure liquid chromatography. Parallel purification product from an antigen-negative cell line failed to demonstrate antibody inhibiting activity. We conclude that greater than 400-fold purification enrichment of antigen can be achieved. We postulate that the partially purified antigenic determinant is a glioma-associated determinant of highly restricted expression and is presented in hapten-carrier form by the glioma cells.

Glioma-associated antigens defined by monoclonal antibodies against an avian sarcoma virus-induced rat astrocytoma

S69-c15 is a highly immunogenic cell line derived from an avian sarcoma virus (ASV)-induced astrocytoma in F-344 rats. Monoclonal antibody (Mab) production was attempted by fusing F-344 rat splenocytes and mouse P3 X 63/Ag8.653 myeloma cells after a syngeneic immunization protocol. 336 fusion clones were screened by cell surface radioimmunoassay (CS-RIA) against the immunizing line S69-c15, rat kidney fibroblast line S203-c11 and Walker rat carcinoma line. Mabs 7G4, 9F1, 10E3 and 10E7 which reacted only with S69-c15 were chosen. Further analysis demonstrated that these Mabs reacted only with rat (13/23 astrocytomas, 2/4 gliomas, 1/11 neurinomas) or mouse (2/10 astrocytomas) neurogenic tumor cells induced by both viral and chemical agents. Reciprocal competition assays suggested that 7G4, 9F1 and 10E3 recognized the same epitope and that 10E7 reacted with a spatially close determinant. Antigen activity could not be found in adult rat tissues (brain, heart, lung, liver, kidney, spleen, thymus, intestine, muscle and peripheral nerve) and fetal brain (8, 12, 20 days gestation) by either absorption analysis or tissue staining. Preliminary characterization indicated that the epitope may be polypeptide-associated. Further antigen purification and tumor localization can be attempted with these Mabs.

Characterization of a chemically homogeneous tumor antigen from a methylcholanthrene-induced sarcoma, Meth A

A tumor antigen isolated from the cytosol of a methylcholanthrene-induced sarcoma (Meth A) has been purified to homogeneity by the criteria of two-dimensional gel analysis and NH2- and COOH-terminal sequencing. The purified antigen has a mol. wt of 82,000 by SDS gel electrophoresis. However, the apparent mol. mass of the antigen was found to be 71,600 and 67,700 by gel filtration chromatography and sedimentation analysis, respectively. It is not a glycoprotein, possesses an acidic isoelectric point (6.0) and exists as dimeric and monomeric species. The dimer is not held together by disulfide bonds. The purified protein retains its ability to induce transplantation immunity in syngeneic hosts when challenged with Meth A sarcomas. Chemical analyses of the NH2- and COOH-termini gave the following sequences: NH2-PKPINVRVTTMDAELEFAIQPN and IDE(F,A)EM-COOH, respectively.

A mouse tumor-specific transplantation antigen is a heat shock-related protein

A tumor-specific transplantation antigen has been purified to homogeneity from the cytosol of a methylcholanthrene-induced tumor, Meth A. The purified antigen is highly immunogenic and specific against challenge with Meth A, providing greater than 95% inhibition of tumor growth in immunized syngeneic mice. Immunofluorescence analysis of Meth A showed that the antigen is a highly abundant cytosolic protein but that it is also present at the cell surface and, therefore, accessible to the host's immune system. The antigen consists of two polypeptide isoforms present in equimolar amounts, having similar masses (84 and 86 kDa), pI values (4.95 and 4.90), and amino acid compositions. Both are phosphoproteins, and neither is glycosylated. The NH2-terminal sequences of the two isoforms are identical except that each chain contains a portion of unique sequence. Comparison of the NH2-terminal and CNBr-fragment sequence data to the sequences of the yeast and Drosophila heat shock proteins (Hsp90 and Hsp83, respectively) reveals that 73 of 91 residues compared are identical. In addition, an anti-Meth A tumor antigen serum that defects the isoforms from a variety of tumors also immunoprecipitates proteins of identical mass and pI from both normal and heat-shocked mouse embryo cells.

The Rauscher-MuLV-induced leukemia, RBL-5, bears two tumor-associated transplantation antigens expressed on distinct molecules

Tumor cells frequently express on their surface a new antigenic determinant which renders them immunogenic in the host animal. When immunity to this antigen results in rejection of a syngeneic tumor transplant, it is referred to as a tumor-associated transplantation antigen (TATA). RBL-5 is a Rauscher murine leukemia virus (MuLV)-induced leukemia of C57B1/6 origin that is potently immunogenic and shares a TATA with other tumors induced by the closely related Friend and Moloney-MuLVs (FMR-TATA). We have recently isolated a 175 kilodalton (kd) glycoprotein (gp175) which has all the properties expected of the FMR-TATA (Rogers et al., 1984). When this molecule was separated from a purified total glycoprotein fraction by DEAE chromatography, the remaining glycoproteins still contained a highly immunogenic TATA. Control experiments involving radioimmunoassay and immunoprecipitation with rabbit anti-gp175 indicated that this immunogenicity was not due to residual gp175 or breakdown products of gp175. We therefore conclude that RBL-5 expresses at least two distinct TATAs: gp175 and another glycoprotein distinguished from gp175 by its elution from a diethylaminoethyl-cellulose (DE52) column. These results, from a completely in vivo system, support data with other tumors obtained by in vitro methods and indicate that tumor cells may express several immunogenic molecules.

Humoral response against murine lymphoma cells xenogenized by drug treatment in vivo

Treatment of murine lymphomas with triazene derivatives may lead to the appearance of novel drug-mediated tumor antigens, a phenomenon known as chemical xenogenization. Such antigens, which are capable of eliciting specific transplantation resistance in histocompatible mice, have been previously detected by in vivo and in vitro cell-mediated immune responses. In the present report we address the question of the humoral antibody response to a chemically xenogenized lymphoma. Histocompatible mice were given several injections of live cells of the xenogenized tumor. Ten days after each immunization, pooled sera from different animals were analyzed for Ab content by means of flow microfluorometry analysis and CEL-ISA assay. The results reveal that antibodies of both IgG and IgM classes capable of binding the xenogenized tumor can already be detected after one single sensitization. However, the Ab titer gradually increases through subsequent immunizations, reaching a peak level after 3-4 injections at a time when most of the humoral response is made up of antibodies of IgG class. The specificity of the anti-xenogenized tumor hyperimmune sera was subsequently investigated by its reaction with the parental, non-xenogenized line and with normal tissue cells of the same or allogeneic haplotypes. The data obtained point out that cross-reactivity with the parental line could be completely removed by absorption of the hyperimmune sera on parental cells, which removed most of the IgM antibodies. Moreover, the presence of an excess of anti-parental antibodies on the xenogenized tumor cells does not prevent the subsequent binding of the hyperimmune absorbed serum, thus indicating that the novel determinant(s) recognized on xenogenized cells are not spatially related to those shared with the original parental tumor. In addition, the hyperimmune absorbed serum does not cross-react with normal hemopoietic or lymphoid cells of the same (H-2d) or allogeneic H-2b and H-2k haplotypes. Furthermore, no alien histocompatibility antigens of H-2b or H-2k haplotypes could be detected on the xenogenized tumor cell surface. Taken together, these data provide evidence that chemical xenogenization of a murine lymphoma leads to the appearance of novel determinant(s) detectable by specific antibodies.

Neoplastic cells as targets of spontaneously cytotoxic lymphocytes: studies with natural killer-like cell lines

Native natural killer (NK) cells comprise a heterogeneous family of lymphocytes distributed among several organs, which display spontaneous cytotoxic reactions directed against a broad range of tumor targets. In these studies, murine cell lines have been established in vitro following the selective expansion of bone marrow- and spleen-derived killer progenitors in culture medium supplemented with interleukin-2. Several clones of independent origin have been characterized in order to determine the extent of their phenotypic and functional diversity. With few exceptions most of them were found to be highly effective in lysing a variety of tumor cell lines, to share common cell surface alloantigens, lectin-binding receptors, and cytochemical markers. The presence of prominent azurophilic cytoplasmic granules is the most characteristic ultrastructural feature of these cells. In attempting to elucidate the nature of membrane components specifically recognized by NK cells we compared several isogenic tumor cell variants selected on the basis of their differential NK susceptibility, immunogenicity, metastatic potential or resistance to cytotoxic plant lectins. Sialylated glycoconjugates exposed on the external face of the tumor cell membrane appear to be essential determinants in the interaction between NK cells and their targets. Permanent cell lines retaining most of the functional attributes of endogenous NK cells may prove instrumental in understanding their role during tumor progression.

Isolated soluble fractions from the murine B16 melanoma induce primary in vitro syngeneic antitumor responses

This paper extends our previous studies, which documented our ability to isolate immunogenic entities from nonimmunogenic or weakly immunogenic tumors. B16 melanoma cells failed, in our in vitro experimental system, to induce anti-B16 cytotoxic responses in spleen cells derived from normal syngeneic C57BL/6 mice. The B16 melanoma cellular homogenate was fractionated on an Ultrogel AcA 34 column, and the various fractions were tested for their ability to induce anti-B16 cytotoxic responses under the same conditions as those used for intact B16, the nonimmungenic tumor cells. Certain fractions, some of them with relatively low protein concentrations, induced anti-B16 cytotoxic responses in spleen cells of normal C57BL/6 mice, whereas others, some of them with relatively high protein concentrations, failed to induce such responses. One fraction (Fr.), designated Fr. 5/6, was examined in detail. It was found that in normal syngeneic spleen cells this fraction induced effector cells that efficiently killed (at various E : T ratios) the relevant B16 target cells and RBL5 syngeneic tumor cells, but not the YAC allogeneic tumor cells or C57BL/6 lymphoblasts. Furthermore, an excess of unlabeled B16 cells most efficiently blocked the ability of these anti-B16 effector cells to kill radiolabeled B16 target cells. RBL5 tumor cells, YAC tumor cells, or C57BL/6 lymphoblasts failed to block these effector cells efficiently. A significant fraction of the effector cells induced with Fr. 5/6 was characterized as thymus-derived cells (Thy-1+, Ly-2+3+ cells). It was suggested that another fraction of the cellular population was natural killer cells, which cytolyzed the RBL5 target cells. Various theoretical and practical aspects of these findings are discussed.

Carrier and adjuvant properties of liposome-borne tumor-specific antigens

The purpose of this study was to examine (1) the association of tumor extract proteins with phospholipid vesicles of varying physiochemical properties, and (2) the adjuvant and carrier properties of liposome-borne tumor antigens in the in vivo induction of an antitumor immune response. Cell surface antigens of the 3-methylcholanthrene-induced fibrosarcoma of C3H/HeJ mice, MCA-F, were extracted using 2.5% 1-butanol. Crude and electrofocused antigen preparations capable of eliciting a protective antitumor immune response were used to prepare liposome vaccines. The incorporation of extract proteins into liposomes formed by butanol dialysis (BVD) was three- to five-fold greater than the encapsulation of protein into the aqueous compartment of multilamellar vesicles (MLV). The electrochemical properties of the BDV had a significant effect on the induction of an antitumor response: Antigens borne on negatively charged, but not uncharged, liposomes were effective in protecting hosts against supralethal tumor challenge, and displayed a specific activity 20- to 50-fold greater than soluble antigen. Antigens carried by MLV were not effective in generating an immunoprotective response. The lipophilic characteristics of butanol-extracted antigens allowed (1) the passive adsorption of immunoprotective tumor antigen onto the surface of preformed vesicles, and (2) adsorption of MCA-F antigen onto the surface of an antigenically distinct tumor MCA-D. In the latter experiment, adsorption of MCA-F-specific antigen onto MCA-D cells resulted in a change in the membrane antigen phenotype as measured by indirect immunofluorescence. Although butanol released a lipophilic moiety from cells which spontaneously reassociated with phospholipid bilayers, no evidence for a lipoidal antigen was obtained when tumor-derived lipids were used as immunogens. This study demonstrates that butanol-extracted tumor antigen is lipophilic without being a lipid, and that negatively charged liposomes can be effective as carriers and adjuvants for tumor antigens in the induction of an antitumor immune response.

Purification and biochemical properties of tumor-associated transplantation antigens from methylcholanthrene-induced murine sarcomas

A tumor-associated transplantation antigen with an apparent molecular weight of 75,000 has been isolated from the cytosol of the BALB/c methylcholanthrene-induced sarcoma, Meth A. The antigen was purified either by preparative electrophoresis in the presence of NaDodSO4 or by immunoaffinity chromatography after hexylamine agarose chromatography, gel filtration, and hydroxylapatite chromatography. The 75-kilodalton (kDal) protein prepared by either of these methods effectively primed BALB/c mice to reject the Meth A tumor; such priming provided no protection against challenge by other independently derived sarcomas of BALB/c origin. A second protein, also 75 kDal, was isolated from the cytosol of the recently derived methylcholanthrene-induced sarcoma CI-4 by essentially the same chromatographic scheme. This protein also was immunogenic in the tumor rejection assay and provided protection only against CI-4 challenge. The antigens purified from the Meth A and CI-4 sarcomas appear to be closely related proteins. Both of them can be purified from the cytosol fraction and can be recognized by a rabbit antiserum prepared against the Meth A 75-kDal protein. The two proteins have approximately the same molecular weight, have similar but not identical amino acid compositions, and differ in their chromatographic behavior on hexylamine agarose and hydroxylapatite as well as in their isoelectric points. These results indicate that the individually specific transplantation antigens found in chemically induced sarcomas may be the products of a single multigene family or somatic derivatives of a single gene.

Immunogenicity of cellular and acellular antigen preparations from a methylcholanthrene-induced mouse sarcoma

Inbred mice were immunized with various cellular and subcellular preparations from a syngeneic, methylcholanthrene-induced sarcoma to induce transplantation immunity against a subsequent challenge of viable cells from the same tumor. Only with conventional immunization procedures using radiation-attenuated tumor cells or viable tumor cells in admixture with bacillus Calmette-Guérin was transplantation immunity obtained. Neither native soluble tumor extracts prepared using several extraction methods, nor cross-linked preparations of these extracts gave rise to transplantation immunity. With few exceptions, however, the immunizing preparations resulted in the production of tumor-specific humoral antibody. The manner in which tumor antigens are presented to the host as well as the experimental system (animal strain; tumor) used appear to play an important role in the development of transplantation immunity against tumors.

Purified membrane fractions from mammary tumor cells elicit biological reactivity in in vitro cell-mediated immune reactions

Lithium chloride (LiCl) aggregation followed by centrifugation through a sucrose step gradient was used to obtain purified plasma membranes from two sublines of mouse mammary adenocarcinoma. These two tumor lines were chemically induced by treatment of a hyperplastic nodule with 7,12-dimethylbenzanthracene (DMBA). One line, D1-DMBA-3, has been found to be immunogenic to the host of origin, while the other, D1-DMBA-2, does not elicit specific tumor immunity as previously tested in in vivo and in vitro immune reactions. The purified membrane fractions were assayed for protein, DNA and sialic acid content as well as enzymatic markers of membrane purity. When tested in blastogenesis and cytotoxicity reactions, membrane-containing fractions from the D1-DMBA-3 immunogenic tumor were found to be stimulatory to spleen cells of D1-DMBA-3 tumor bearers over a ten-fold protein concentration. Spleen cells from normal mice do not respond in these reactions to the various tumor fractions. No reactivity was observed when non-membrane-containing preparations were used as stimuli in the cell-mediated immune reactions. The specificity of these reactivities was further demonstrated by the lack of responses when fractions from the non-immunogenic D1-DMBA-2 tumor were tested in parallel in our in vitro assays. The data presented indicate that the procedure employed is useful for the isolation of membrane-associated, tumor-specific antigens which can be easily quantitated and still retain biological activity in in vitro tests of cell-mediated immunity.

Modification of cell cholesterol content of rat tumour cells. Effect upon their tumorigenicity and immunogenicity

Among the different constituents of the cell membrane, lipids have been poorly studied with respect to their role in the immunogenicity of tumour cells and their influence on the expression on tumour-associated antigens. Since liposome-associated antigens are more potent immunogens when the lipid matrix is in a rigid state, we have modified the lipid composition of rat hepatoma cells by incorporation of cholesteryl hemisuccinate (CH) into the lipid matrix, and studied its effect on the tumorigenicity and immunogenicity of these tumour cells in syngeneic animals. A slight and significant decrease of tumorigenicity of CH-enriched D23 cells was observed when 2×103 cells were injected SC, whereas with a higher tumour cell challenge there was no difference in the tumorigenicity of untreated or treated cells. The immunogenicity of CH-treated cells was tested by IP immunization with 107 or 106 cells followed 1 week later by an SC challenge with 2×104 viable D23 cells. No statistical difference was observed between the immunogenicity of CH-enriched cells and that of control cells on either tumour incidence or tumour growth rate. In addition, similar experiments performed with the spontaneous mammary carcinoma SP4 showed that CH-enriched SP4 cells were of lower immunogenicity and unable to induce a significant memory immunity. This lack of effect of the CH treatment on the immunogenicity was not related to the absence of incorporation of CH, since the CH treatment increased the cell lipid rigidity as determined by the increase of fluorescence anisotropy of the diphenyl hexatriene probe. These results obtained in two weak immunogenic tumour models underlined the need for further studies before such a lipid modification of cancer cells is applied in human immunotherapy trials.

Antigenic variation in cancer metastasis: immune escape versus immune control

Antigenic variation in cancer metastasis was observed in a syngeneic murine tumor system consisting of a low metastatic parental tumor line (derived from a methylcholanthrene-induced DBA/2 T lymphoma, Eb), a high metastatic spontaneous variant thereof (ESb), and a low metastatic 'revertant' from ESb (ESb-M). All three lines expressed tumor-associated transplantation antigens (TATA) which elicited specific T cell-mediated antitumor immune reactions in the host. The strongest host response was elicited upon intradermal inoculation. It could be followed by (a) the infiltration of the locally growing tumor by host cells, such as lymphocytes and macrophages, (b) the establishment of specific systemic antitumor immunity, (c) the generation of immune cells capable of transferring protective antitumor immunity into a normal syngeneic recipient, and (d) the generation of tumor specific cytotoxic T lymphocytes (CTL). Anti-TATA CTL were used as typing reagents to investigate the stability or variability in the TATA expression by cloned tumor cell lines. Antigenic variability in the TATA expression was seen under various conditions: (a) clone-dependent variation in the sensitivity to anti-TATA CTL lysis upon prolonged growth in tissue culture, (b) qualitative change in the TATA (TATA1 leads to TATA2) upon successive i.p. transplantation of the parental Eb tumor line and, (c) generation of TATA negative immune escape variants (TATA2 leads to TATA-) during metastasis formation from a s.c. site. The relative inefficiency of specific immunization procedures against ESb was found to be due to the effective generation of TATA negative variants by this highly metastatic tumor. The balance between immune control and immune escape could be influenced to the advantage of the host by some means, for instance optimizing the route of antitumor-immune sensitization or by infusion of allogeneic but H-2 identical antitumor-immune T cells. Such immune cells recognized the tumor via minor histocompatibility antigens and thus circumvented the need of TATA recognition. Finally, manipulations at the cell surface of the highly malignant ESb tumor such as those introduced in the ESb-M variant were found to dramatically effect its metastatic potential.

Non-immunogenicity of enucleated rat hepatoma cells in syngeneic animals

Cytoplasts and karyoplasts were obtained by ultracentrifugation of Hepatoma D23 cells on a Ficoll gradient containing cytochalasin B. Their nuclear and protein content and their metabolic activity were determined. Three i.p. injections of 2.3 x 10(7) cytoplasts were unable to protect syngeneic WAB/Not rats against an s.c. challenge of 10(4) D23 cells, whereas a similar amount of karyoplasts, or 3 injections of 10(6) irradiated D23 cells, were fully protective. Ability of cytoplasts to act as primary or secondary immunogen were also studied, and compared to that of 0.01% glutaraldehyde-treated cells, 43 degrees C heat-treated cells and 3M KCl-soluble extracts, these preparations also being of weak immunogenicity. Only heat-treated cells behaved as a primary immunogen, whereas none of the preparations provided a secondary stimulation. Moreover, when these preparations were fed in vitro to peritoneal-exudate cells before their injection into rats, cytoplasts and glutaraldehyde-treated cells showed no immunogenicity, whereas heat-treated cells induced full protection against tumour challenge. Therefore, in this tumour model, the in vivo persistence of immunogen and the presence of a nucleus are likely to be crucial in inducing transplantation resistance to tumour.